Modelling the concentrations of dissolved contaminants (Cd, Cu, Ni, Pb, Zn) in floodplain soils.

Abstract

Central European floodplain soils are often contaminated with potentially toxic metals. The prediction of their aqueous concentrations is a prerequisite for an assessment of environmental concerns. We tested the aqueous concentrations of cadmium (Cd), copper (Cu), nickel (Ni), lead (Pb) and zinc (Zn) derived from multi-surface adsorption modelling (on hydrous iron, aluminum and manganese oxides, clay and soil organic matter) against those analyzed in situ in the soil solution of four horizons of floodplain soils at the Elbe River, Germany. The input data for the reactive metals were derived from a seven-step sequential extraction scheme or from extraction with 0.43M nitric acid (HNO3) and evaluated in four modelling scenarios. In all scenarios, measured and modelled concentrations were positively related, except partially for Pb. Close reproduction of the measured data was obtained using measured data of accompanying cations and anions together with amounts of reactive metals from both the sequential extraction or from 0.43M HNO3 extraction, except for Cu, which was often strongly overestimated, and partially Cd. We recommend extraction with 0.43M HNO3 to quantify reactive metals in soil because the modelling results were metal-specific with better or equal results using the single extractant, the application of which is also less laborious. Approximations of ion concentrations and water contents yielded similar results. Modelled solid-phase speciation of metals varied with pH and differed from that from sequential extraction. Multi-surface modelling may be an effective tool to predict both aqueous concentrations and solid-phase speciation of metals in soil.